This was read by an AI using a sample of my actual voice.
Turns out: When your sign up flow consists of users entering an email address and password you might want to verify that the person signing up actually owns that email address. Who would have guessed.
Since I’m using Firebase to build Goblin Mode I have a whole plethora of features and functions at my disposal that make this easily possible. In theory.
In practice I would have had to learn Firebase cloud functions to delete accounts without a verified address after a specific timeframe and that just wasn’t realistic. I’ll teach myself how cloud functions work when I start working on notifications.
Well, how to solve the issue of people being able to sign up with other people’s email addresses? You circumvent the whole thing by not offering email/password sign up at all.
Goblin Mode sign ups now use Sign in with Apple. Ethically it’s the only right choice if I want to even think about releasing the thing at some point. I’m not good enough at programming yet to say with conviction that I could build a confidence-inspiring login system any other way.
But that’s fine. This flow is surprisingly sleek and Apple even offers to hide people’s actual email addresses.
Oh and I used this opportunity to slightly tweak the whole onboarding experience. It’s not my greatest design work ever but it does the job.
A lot has happened since I shared thesetwo posts about the project I’m currently working on. I figured it’s time to let you know what has happened in the meantime. Here’s the short version: I learned a lot. Like A LOT. This is my third app ever and it’s the one where a lot of previously vague concepts finally clicked. I’m injecting dependencies, juggle with types and create view models like there’s no tomorrow. It’s so much fun!
What was supposed to be a placeholder icon has grown on me over time
A couple of weeks ago I invited some friends to try out the app and see if the overall idea works. This was a huge moment for me. People signed up for something I programmed. They created literal user accounts saved to a database I connected to an app I coded. Crazy. It got even cooler after that: They started using the app. And somehow they didn’t stop. I’m not sure if they’re only using the App because they want to be nice but it feels like they enjoy sharing one-off status updates with a closed and private group of friends. No endless timeline, no fame, only letting people know what you’re up to.
Let me give you an overview of how the app looks like at the moment.
This is the main view. You see your own current status and how old it is. Below that you’ll see all current status of your friends.
I wrote the code for all of this like three separate times. The first two iterations didn’t care about the amount of database requests at all. They worked but they weren’t scalable. I’m now using Firestore realtime updates and it’s amazing. If a friend updates their status while you’re in the app, the status automatically gets loaded from the server and displayed on top of the stack.
But what happens when a friend hasn’t updated their status since midnight?
SPIDERS!
They become stale. I like the idea of having a predefined tabula rasa moment. Each day starts with a clean slate. If you post something one minute before midnight that status will only be fresh for one minute. Slightly weird, but I kind of like it.
Speaking of likes: I implemented a system to like status. It works well and is fun. I’ll probably replace it with a system to react with a tag, though. Liking feels wrong when a friend posts a status about them being tired, feeling empty or being sad. I want to be able to react to these with “🤡 Clown”.
There are more than 130 tags now. I built an admin area and support for admin accounts into the app and can CRUD tags and tag categories in seconds. New tags and categories show up on user’s devices without the need for an app update. It’s all in the cloud, baby!
I also build a system for tags and tag categories that are behind the paywall. My current idea for monetizing the app would be to put most of the cool tags behind a subscription of some kind. Those server bills need to be paid!
With the ever increasing amount of tags my friends asked for a tag search feature. That has been implemented as well. For those moments when you need to share your shower thought quickly.
And last but not least: I revamped the whole follow/unfollow system I talked about here. It became clear that asynchronous following isn’t the right concept for an app that’s meant for you to share semi-private status updates. Now there’s a whole screen dedicated to see who requests being friends with you. If you accept a request you become friends with the other person and you can both see the status of each other. If you cancel a friendship both ex-friends stop being able to see the respective status. Fair and easy.
I’m toying with the idea of limiting the maximum number of friends each person can have. Just to drive the point of the app home. That didn’t work out very well for Path back in the day but maybe times have changed. Who knows.
That’s it for now. I still don’t know if I’m going to release this to the public at any point. I’m still a novice coder and my implementation of everything related to Firebase has a big potential to suck. I’m generally cool with my code not being the best but in this case it could result in me racking up an enormous Firebase bill. That would not be fun at all.
As I said: Who knows. I might find a way to limit the app features enough to let people try it out without me becoming poor while allowing people to pay to cover the database costs. It sure would be an interesting problem to have.
Do you feel like there’s a Goblin Mode shaped hole in the lives of you and your friends?
I’m having the time of my life over here. Without getting too much into what my next app will be: Look at this! All the data gets fetched from Firebase. Creating a new tag uploads it to firebase. The categories are documents in Firebase as well. The view showing categories and tags together is based on an array of objects that combines both. I’m juggling with data!
I wouldn’t have been able to do any of this a couple of months ago. We’re looking at a logged in user, authenticated through email and password, that has an isAdmin flag set to true. Only admin users are able to see the tag management section and are able to create new ones.
Databases! Optionals! ViewModels! Aysnc/Await! Everything comes together.
This is so much fun that I have to force myself to stop working to get a good night’s sleep.
Some of you might have guessed it: I’m a big fan of the internet. The simple fact that I could work through this amazing seven(!) hour video about SwiftUI and Firebase by Stephan Dowless for free is still mind-blowing to me. It is free knowledge. Potentially life-changing information just lying around on the internet and all you have to do is consume it.
Stephan is an incredibly enthusiastic instructor who made the time fly by. Even though he obviously doesn’t do this for the first time, he still bursts out stuff like “How fucking slick is that!?” all the time. I felt thoroughly entertained. This is easily one of the most helpful courses I’ve worked through.
I know now that Firebase is the way to go for what I’m currently looking for. The course went through the whole process of building a low-tech Twitter clone. From signing up to fetching data from Firebase to storing images and writing and liking tweets. Not only do I now understand the basics of Firebase (and how to work with databases in general), I also got a better insight into how slightly more complex apps are supposed to be structured. MVVM is easier to understand when applied to a project with a certain complexity. I didn’t get it when other people tried to explain it with simpler examples.
This was very well invested time and I feel like I’m ready to start my next project. Something with databases! On the internet! How exciting!
Today I worked through this tutorial and came away with the conclusion that Firebase seems to be exactly what I was looking for. The Firebase SDK seems to be straight-forward enough to work well with my current skillset.
This code here creates a user that can be authenticated by email and password. Just like that. I don’t have the slightest idea what’s happening behind the scenes but to be honest: At this point I don’t really care. I’m just happy that it works.
func register() {
Auth.auth().createUser(withEmail: email, password: password) { result, error in
if error != nil {
print(error!.localizedDescription)
}
}
}
After working through the tutorial I even managed to implement a couple of experimental test functions myself. I’m off to a good start and managed to recreate my motivation by approaching the problem from a different angle.
After spending a couple hours with tutorials and courses about CloudKit, I’m a little deflated. What I found was either outdated or at least not current and while I’m also trying to get an understanding of previous solutions, I want to concentrate on learning what’s going to be the standard of the future.
I kind of hit a wall with CloudKit there. The one very good course I worked through uses Combine but CloudKit offers methods for Swift’s “new” async/await as well. My knowledge of Combine is now larger than ever but I don’t want to go ahead and start building apps with something I feel like will be looked at as outdated code very soon.
There’s hope though. Because I don’t have anyone to pester with questions, I have to rely on good online content. That’s why I’m going to have a look at Firebase. A user friendly SDK and a shitton of noob friendly tutorials might be a better experience for my database needs for now.
